Doing the Math: Testing Capture Odds
TL:DR – The visible percentage shown is “the odds of a capture, starting now”. It’s the most intuitive interpretation, and experimentally checks out.
So, I did the thing (and will continue doing more after today’s silly real life obligations). I made hundreds of balls, and recorded the shown percentage (scope, first wiggle, second wiggle) of every single throw. Then calculated the actual percentages vs those shown in game.
The result was the intuitive answer: the number you see is the chances that the Pal will be captured. Looking through a scope and see “13%”? If you throw a ball, there’s a 13% chance it will be captured. (NOT a 13% chance of deflection, or anything like that). This means there’s an 87% chance that it will either: deflect the ball, break out on first wiggle, or break out on second wiggle; and a 13% chance that the capture will go through completely.
Similarly, the percentage shown after initial grab, and after first wiggle, are the odds that the capture will finish successfully from there, not the odds of the specific check that’s being made.
The actual checks being made for a 13% throw are experimentally close to: 42% chance of grab, 53% chance of holding first wiggle, 64% chance of holding second wiggle. The combined odds of those 3, is what shows as 13% (in my experiment, closer to 14.2% real odds, but within a reasonable confidence interval given the sample size)
There are a handful more cases I want to test, such as what the net effect from lifmunk effigies really is, or how the back-capture affects things or better balls – but i’m out of time for this morning. I’ll come back and update once I have the chance to do more testing!
There may be bugs and non-intuitive things as you move up in higher levels and higher balls etc. There’s definitely some rounding going on (eg: 13% may be actually 13.4%, or 12.8% etc) but without mining the actual code for how the tables are being calculated the best I can do is say that they’re at least accurately very close to the real odds.
So when someone pulls a 0.03% capture, it’s really a 3 in 10,000 chance (or close to it), which really isn’t all that surprising given how many balls are being thrown every minute across the game.
That makes a lot of sense. I see it less 13 chance and more probability. Like you need three yes conditions normally. Each with there own odds. probability you get three yes’s is 13% and 87% you get one no. Would explain why I didn’t see a change when trying to catch low level pals that I hadn’t aggro’d yet. Considering I got a message to lower health after a deflect I think health, and status only affect deflect chance. Since deflect is first those have the biggest impact on scope probability. Wiggles base is probably mob type and level vs player lvl. Also why it seems to have a low affect. Like if a 10 CP is 10% increase over base on each wiggle. That would only be 1% probability for success on both wiggles. It would only increase the overall probability by .42% in your example. Imo if not already the CP should be a square root curve based on how far base probability is from 100%. So on harder to catch mobs you have more incentive to use better spheres but are discouraged from using on weaker mobs. Would also improve effigy impact which I think is to low
I do not mean to stack more work upon you. Though I have had 100% captures break out twice and be caught by the third ball. How did upgrading your effigy factor into this as well? If I aim and it says 50% and I hit and it says 78% and after first wiggle it says 90% sometimes it still takes 10-15 balls to capture.
To what with my limited math skills doesn’t make sense unless it isn’t factoring in the boosted capture rate.